Some weird things are happening on our machine related to utmp entries.
We're running DU3.2C.
Seems that talk requests and biff notices are going to the wrong people.
Investigation shows that there are multiple entries in utmp for the same
tty. One is marked DEAD_PROCESS and one USER_PROCESS. For some reason biff
is choosing the DEAD_PROCESS entry.
Any ideas? We _do_ run screen 3.7.1 which may be affecting things. I've got
it compiled to always log screens in which means that the initial login
gets removed until the screen gets detached. Neither of the 2 users in this
case has any screen processes running.
These events are not limited to particular sets of users and it's happening
to at least biff/comsat and talk.
utmp is mode 644 owned root.adm
--
[=======================================================================]
[ Kevin Lentin |finger kevinl_at_fangorn.cs.monash.edu.au| ]
[ K.Lentin_at_cs.monash.edu.au |for PGP public key block. Fingerprint | ]
[ Macintrash: 'Just say NO!' |6024308DE1F84314 811B511DBA6FD596 | ]
[=======================================================================]
Received on Wed Feb 14 1996 - 05:44:08 NZDT